Anda di halaman 1dari 11

ARTICLE IN PRESS

European Management Journal (2014)

Contents lists available at ScienceDirect

European Management Journal


j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / e m j

Factors inuencing popularity of branded content in Facebook


fan pages
Ferran Sabate a,*, Jasmina Berbegal-Mirabent b, Antonio Caabate c, Philipp R. Lebherz d
a
b
c
d

Universitat Politcnica de Catalunya BarcelonaTech, Department of Management. C. Jordi Girona, 1-3, Edici C5. 08034, Barcelona, Spain
Universitat Internacional de Catalunya, Department of Economy and Business Organization. C. Inmaculada, 22, 08017, Barcelona, Spain
Universitat Politcnica de Catalunya BarcelonaTech, Department of Management. C. Jordi Girona, 1-3, Edici C5. 08034, Barcelona, Spain
Karlsruher Institut fr Technologie (KIT), Kaiserstrae 12. 76131, Karlsruhe, Germany

A R T I C L E

I N F O

Article history:
Received 22 October 2013
Accepted 11 May 2014
Available online
Keywords:
Content marketing
Social networking site
Consumer engagement
Social media optimization
Facebook brand page

A B S T R A C T

Social media is achieving an increasing importance as a channel for gathering information about products and services. Brands are developing its presence in social networking sites to meet brand awareness, engagement and word of mouth. In this context, the analysis of the factors that are conditioning
consumer interaction with branded content becomes a matter of interest. This paper aims to shed light
on those factors that are expected to impact on Facebook branded post popularity. A conceptual model
is developed to reect the inuence of the contents richness and time frame on the number of comments and likes. An empirical analysis using multiple linear regressions is conducted based on 164 Facebook posts gathered from the fan pages of 5 Spanish travel agencies. Results suggest that the richness of
the content (inclusions of images and videos) raises the impact of the post in terms of likes. On the other
hand, using images and a proper publication time are signicantly inuencing the number of comments, whereas the use of links may decrease this metric.
This study empirically contributes to the existing literature on the management of marketing strategies for consumer engagement in social networking sites.
2014 Elsevier Ltd. All rights reserved.

1 Introduction
Social networking sites (henceforth, SNS) have become very
popular and have been increasingly attracting the interest of marketers. They account about 6% of all website visits done and 19% of
all time spent online (Radwanick, Lipsman, & Aquino, 2011; Tuten,
2008). Social media is achieving more and more importance as a
channel for gathering information about products and services and
to take prot of new opportunities (Verhoef & Lemon, 2013).
SNS have stimulated new ways of interacting, shaping new forms
in which people communicate, make decisions, socialize, collaborate, learn, entertain themselves, interact with each other or even
do their shopping (Constantinides & Fountain, 2008; Hanna, Rohm,
& Crittenden, 2011; Hansen, Schneiderman, & Smith, 2011; Mangold
& Faulds, 2009). Consequently, the study of social media and its
effects on consumers and organizations is increasingly attracting academic attention while it opens new research avenues for strategists and marketers (Bughin & Manyika, 2009; Constantinides &
Fountain, 2008; Fischer & Reuber, 2011; Urban, 2003).

* Corresponding author. Tel.: +34 93 401 56 31.


E-mail address: ferran.sabate@upc.edu (F. Sabate).

Among others, benets arising from a well-designed social media


marketing strategy may materialize in a better grasp of consumers behaviors and preferences, making consumers share the brands
message as word of mouth to their peers, connecting to consumer
for improvement and R&D processes, increasing brand engagement and brand message exposure, as well as driving trac to corporate websites (Hettler, 2010; Smith & Zook, 2011; Tuten, 2008).
People are gradually shifting their trust to recommendations and
experiences from other consumers. SNS allow users to publish and
interchange opinions and experiences about brands and their products and services. Several studies conrm the inuence of this user
generated content on purchase intention, and that this inuence
applies for different kinds of products and services (Chevalier &
Mayzlin, 2006; Dhar & Chang, 2009; Duan, Gu, & Whinston, 2008;
Rehmani & Khan, 2011; Sierra Snchez, 2012; Ye, Law, & Gu, 2009).
This behavior, known under the term of word-of-mouth (Brown,
Broderick, & Lee, 2007; Kozinets, de Valck, Wojnicki, & Wilner, 2010;
Moldovan, Goldenberg, & Chattopadhyay, 2011; Trusov, Bucklin, &
Pauwels, 2009), contrasts with more traditional marketing communications and seems to be gaining importance with the rise of
Web 2.0. Social media marketing should be oriented rst, to understand clients; second, to create custom-made online content; and
third, to dene a tting strategy in a way that strengthens the reputation of the brand (Hettler, 2010; Heymann-Reder, 2011; Kilian

http://dx.doi.org/10.1016/j.emj.2014.05.001
0263-2373/ 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
2

F. Sabate et al./European Management Journal (2014)

& Langner, 2010). It is therefore a matter of interest to know which


characteristics should have this online content in order to be spread
by consumers without any enforcement or cost, just because people
like to share the content to their peers in a fashionable but direct
and personalized atmosphere (Agresta & Bough, 2011; Constantinides
& Fountain, 2008; Fournier & Avery, 2011; Hettler, 2010).
Content is the instrument that stimulates interaction. Brands must
publish pieces of content trying to address customers motivations delivering interesting content for them when and where
needed. Successful content is adopted by customers, adding value
by sharing it and producing derivative or original content that spreads
through peer-to-peer interactions. Thus, content reaches popularity, thanks to those customers who positively interact with it, contributing to its spreading and becoming brand advocates who can
inuence purchase decisions of others. Postmodern consumers bring
the challenge of addressing these incentives individually and collectively (Sashi, 2012; Scott, 2007; Simmons, 2008; Smith & Zook,
2011).
Within Facebook, dissemination of branded content or posts can
be achieved through several mechanisms. Users who are fans of the
brand will see in their walls this branded content, and then they
can interact with it by liking, sharing or commenting. Each of these
actions potentially promotes the content to all the customers friends
walls. Consequently friends of fans can also contribute to exponentially disseminate this content.
Like most SNS, Facebook allows brands to create proles and
interact with users. Fan pages are brand oriented proles that provide
additional functionalities like detailed analytics and better content
and fans administration. Facebook characteristics provide unique
and interesting conditions for investigating the interaction of multiple selves and the incorporation of brands in consumer selfexpression (Hollenbeck & Kaikati, 2012, p. 396). From brands point
of view and according to the classication of Dholakia, Bagozzi,
and Pearo (2004), Facebook combines characteristics of smallgroup communities, based on pre-existing oine relationships, with
those of network-based communities where a member without
pre-existing relationships connects around the brand through its
fan page. As members of a brand community, participation incentives are based on personal benets. Nevertheless, fans can
also act as brand evangelists spreading branded content through
their friends network where motivations to interact came from
social benets. This double aliation, to personal and brand communities, is essential to understand word-of-mouth and the paths
through which the branded content reach popularity and spreads
virally.
Online social communities are identied as one of the key new
media phenomena (Hennig-Thurau et al., 2010) with research implications for the successful management of customer interactions. Specically, how can communities be used for brand
management and how can a brand acquire virtual consumer friends
are some of the research questions that arise. Based on this rationale, the analysis of the characteristics that makes branded content
popular as well as the study of those factors that are conditioning
consumer interaction become a matter of interest for rms, in
order to address their marketing efforts in social media in the correct
direction.
Accordingly, this work contributes to extend the knowledge on
those characteristics that make branded content popular by identifying how richness and time frame of content publication inuence customers interaction. Moreover, Agresta and Bough (2011)
state that there is no simple formula that guides on how to publish
in social media due to diversity of brands goals and sectorial characteristics. Consequently, we focus our analysis on the Spanish travel
agencies sector by carrying out an empirical study of the posts published by ve Spanish travel agencies on their Facebook fan pages
and the users interaction with them.

We nd that the inclusions of images and videos raise the impact


of the post in terms of likes. Likewise, images and publication time
are signicantly inuencing the number of comments, whereas the
use of links decreases this metric. These results draw implications
applicable for companies social media marketing activities which
are interesting for academics as well as for practitioners.
The remainder of the paper is organized as follows. Section 2 presents the theoretical framework. Section 3 describes the sample and
the methodological approach. Empirical results are offered in Section
4. The discussion of the results and their managerial implications
together with the limitations of this study and potential research
avenues are displayed in Section 5. The paper ends with the main
conclusions summarized in Section 6.
2 Theoretical background
2.1 Drivers for brand post popularity
According to Singh, Jain, and Kankanhalli (2011) there are no theoretical frameworks available yet that could be used to analyze why
and how users contribute to social media. Yet, giving a simple
formula that guides on how to publish in social media is not possible due to the particular circumstances of each brand and because
of the very distinct set of goals and possibilities every business has
(Agresta & Bough, 2011).
In this study we categorize content attributes of SNS according
to a simple classication: whether they are qualitative, based on semantic analysis (soft criterion), or whether they are hints that are
proved in a quantitative and empirical way (hard criterion).
The soft criterion considers the semantics and the interpretation of the message behind a post. Both Scott (2007) and Sterne
(2010) argue that before publishing in SNS businesses should adopt
a consumers perspective and publish only those posts that really
provide value-added information for the reader. The works of
Heymann-Reder (2011) and Hettler (2010) corroborate this statement. In their respective studies they found that those posts revealing funny things of the working environment, news affecting
the business or information that may report direct economic benets to the reader are more prone to capture users attention. These
ndings indicate that post category has a signicant effect over the
user interaction and, as such, should be used for planning of the communications strategy (aDigital, 2011; Pletikosa Cvijikj & Michahelles,
2011).
The main problem of the soft criterion is the diculty in capturing and processing relevant data for analysis. On the one hand,
soft criterion requires a careful content analysis of texts, images or
videos. On the other hand, this analysis can be stigmatized as subjective, as it may be dicult to properly discriminate between those
publications that contribute to enhance the brand, from those that
are damaging it. This leads to a very meticulous analysis which, if
made manually, is very tedious and time-consuming.
Elements of SNS that can be quantied without the need of a subjective interpretation process can be considered under the hard criterion. By this approach, it is possible to compute the frequency and
timing a phenomenon takes place. Likewise, it allows assessing the
richness of the content associated to a post, by simply looking at
the content type that complements the text (i.e. a picture, a video
or a link to another website). These quantitative factors are also characteristics of the published content, but are more easily to capture
and process than the former ones. Thus, given the resource constraints in the collection and processing of data, for the purpose of
this study, we have only focused on those characteristics of posts
that respond to the hard criterion. Accordingly, we aim to identify
those structural features of posts that act as drivers for brand post
popularity. The next section presents the theoretical framework used
and the hypotheses that will be tested.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

Fig. 1. Conceptual model.

2.2 Conceptual model


Fig. 1 schematizes the conceptual framework considered, where
the number of likes and comments represent the metrics to evaluate the popularity of wall posts. We argue that richness (the vividness of the content of the post) and time frame (related to time
and date of publication) have a signicant inuence on brand post
popularity (measured by the number of likes and the number of
comments). Furthermore, the model includes two additional variables, controlling for size differences (number of followers of the
Facebook brand page) and the length of the post (see Fig. 1).
2.2.1 Richness
Different media types entail different capacity for immediate feedback. This is so because of the richness (breadth and depth) of a
message (Daft & Lengel, 1986). This idea of richness is also commonly referred to as vividness of the online content (De Vries,
Gensler, & Leeang, 2012; Pletikosa Cvijikj & Michahelles, 2013).
Indeed, previous studies looking at how to enhance positive attitudes towards a website found that the richness of the message may
play a role (Fortin & Dholakia, 2005).
According to a study conducted by Brookes (2010), images receive
22% more engagement than video posts and 54% more than text
posts, but videos receive 27% more engagement than text posts. These
gures point out that both images and videos are superior to textonly posts, but on the whole, images are denitively more compelling and persuasive than videos.
The aforementioned ndings suggest that the richness of the post
leads to a more proactive attitude toward the wall post. Particularly, the inclusion of dynamic animations (videos), contrasting colors
and pictures (images), and interactive links to other websites (links),
may enhance the salience of a brand post. These mechanisms are
found to stimulate different senses that may increase the users propensity to look at the content of the message compared with those
posts with only text. Therefore, post effectiveness may be conditioned by content type.
In contrast to previous studies where a priori judgments about
progressive levels of richness (low, medium and high) are assumed
(Fortin & Dholakia, 2005; Pletikosa Cvijikj & Michahelles, 2013), in
our approach we only differentiate according to content type (images,
videos and links). This way we avoid any potential subjective bias
on how richness is perceived in the users eyes. Thus, we hypothesize that:

H1c. Posts including links are more likely to generate higher levels
of brand post popularity.

2.2.2 Time frame


In a domain like Facebook, where users prole walls are constantly overloaded with content coming from multiple sources,
posting time is a relevant aspect that should be taken into account
when designing marketing strategies.
Concerning the day of publication, previous studies reveal that
most of the user activities on Facebook are undertaken during
working days (Golder, Wilkinson, & Huberman, 2007). The Buddy
Media Inc. (2011) study also reports that approximately 86% of all
brand postings are done from Monday through Friday, and that customer engagement rates on Thursday and Friday are 18% higher than
on other days of the week. Similarly, the study of Rutz and Bucklin
(2011) on online advertisement supports this thesis, stating that the
click-through rate substantially decreases over the weekend. Therefore, we propose:
H2a. Posts created on weekdays may cause higher levels of brand
post popularity.

H1a. Posts including images are more likely to generate higher levels
of brand post popularity.

It is also suggested that the effectiveness of a post is also inuenced by the time of the day it is published. Identifying the customers habits (peak hours of activity) when determining the posting
schedule is crucial.
According to Golder et al. (2007), users interaction increases
towards the evening, maintaining a steady high level during the night.
Regarding when the post is published, Buddy Media Inc. (2011) found
that brands that post early morning and late at night had engagement rates approximately 20% higher than the average (and that 60%
of all brand posts are done during core business hours (from 10am
to 4pm). Nevertheless, Pletikosa Cvijikj and Michahelles (2013) argue
that if posts are created during those periods with low user activity, when fans will connect (in peak hours) the brand post will appear
at the top of the wall, therefore, the probability for being liked or
commented is higher.
As observed, different temporal patterns are proposed,
however, there is no clear agreement. Also, the sector in which
the rm operates has its own rhythms and, as such, should be
taken into account. Considering this lack of consensus, we decided
to look at our data, and observed that most of the posts were
published within business hours. Aiming at contrasting whether
posts published in business hours are more effective, we
formulate:

H1b. Posts including videos are more likely to generate higher levels
of brand post popularity.

H2b. Posts created during business hours may result in higher levels
of brand post popularity.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

2.2.3 Control variables


Previous research on advertising effectiveness indicates that
message length may affect performance measures such as clickthrough rates (Baltas, 2003). Similarly, the Buddy Media Inc. (2011)
report shows that posts with 80 characters or less have a 27% higher
engagement rate. We therefore include message length as a control
variable, expecting a similar negative effect.
Several studies (e.g. Hong, Dan, & Davison, 2011; Suh, Hong, Pirolli,
& Chi, 2010) have pointed out that social features, such as the number
of friends, followers or similar, have an effect on retweetability and
commenting activity. Particularly, Sun, Rosenn, Marlow, and Lento
(2009) observed that diffusion of wall posts in Facebook reaches up
to 82 levels, signaling that in comparison to real world content, Facebook is capable to spread it faster and involve much more people.
Likewise, in the study of Zhang et al. (2014) on post popularity in
one of the most popular microblogging sites in China, the authors
found that the more followers a user has, the greater potential audience messages posted by this user will have. Aiming at controlling this effect, the number of followers has also been included as
a control variable.

we focused on the evolution of posts published during the previous month, that is, from February to 21 to March 21, 2011. This delay
was necessary in order to capture how users interact with the content
already published. We believe that the time span considered is
enough for the purpose of this research. SNS are characterized for
being extremely fast and dynamic communication channels; hence,
a content posted on the net for more than 30 days is not likely to
receive more interaction.
For the selected period of time 164 posts published by the ve
travel agencies considered were obtained and manually processed. Content shared by other users on these rms fan pages was
not considered.
3.3 Variables

To test the abovementioned hypotheses, we have focused the


study on Spanish travel agencies with a Facebook fan page. The rationale for the scope considered is threefold. First, according to the
report published by Silverman (2012), the travel agency sector is one
of the 10 industries generating major Internet ad revenues. This sector
has also suffered a strong transformation due to the Internet revolution (Buhalis & Law, 2008) and is assumed to be one of the most
actively involved in the use of social media channels (Xiang & Gretzel,
2010). Second, we decided to focus on Spain for two main reasons.
On the one hand, the aforementioned growing importance of social
media channels in the tourism industry also applies for Spain (IAB
Spain Research & Elogia, 2011; Sabate, Canabate, Velarde-Iturralde,
& Grinon-Barcelo, 2010). On the other hand, this country is ranked
in the top ten Tourism Competitiveness Index 2011 published by
Blanke and Chiesa (2012). These two arguments reinforce the interest for studying the Spanish case. Finally, we chose the Facebook platform as it is the largest and most used SNS in Spain (IAB
Spain Research & Elogia, 2011). Moreover, previous literature examining customer engagement in Facebook reinforces the suitability of this SNS (De Vries et al., 2012; Dholakia & Durham, 2010; Smith,
Fischer, & Yongjian, 2012).
Firms were selected according to three different criteria. As we
were interested in travel agencies actively involved in the use of SNS,
a rst criterion considered the number of social media channels in
which the rms have presence. Specically, we checked their presence on Facebook, Twitter, YouTube and Blogs, as these are the most
recurrent social networks (Lawrence, Pownal, Jrg, & Carmo, 2011).
Second, and similar to previous studies (Sabate, Berbegal,
Consolacin, & Caabate, 2009; Sabate et al., 2010), we used the Alexa
trac rank (http://www.alexa.com) to measure the popularity in
terms of visits to the rms website. Third, we included an economic dimension, accounting for the revenues obtained during 2008 and
2009, looking for rms with an important weight in the travel agencys sector.
Five Spanish travel agencies were nally chosen: Rumbo.es,
Atrapalo.com, eDreams.es, MuchoViaje.com and Barceloviajes.com.

Fig. 2 illustrates the information (variables) gathered for each post


on travel agencies fan pages. Detailed description of these variables is provided in Table 1.
Independent variables are those identied with numbers from
1 to 6 (Table 1). Following the classication explained in Section 2,
these variables respond to the hard criterion as they are structural
characteristics related to the posts, rather than capturing the meaning
of the content itself.
In order to t with a linear regression model, variables have been
codied as dichotomous or dummy ones. Codication was really intuitive for links, images and videos as they clearly respond to a yes/
no question when asking for their presence in a post. We decided
not to measure them as a number of items after realizing that the
maximum number of images and videos per post was 1, and that
for the specic case of links, only three observations include more
than 1 (2 links).
Concerning the variable that captures the time of the post (Time),
we decided to differentiate between those posts published during
business hours than those published before/after this schedule (see
Table 1). Although we are aware that working hours may vary from
one rm to another, we believe that the two segments considered
are representative enough of the traditional Spanish workday. Yet,
considering more than two categories would have led to a reduced
number of observations for each segment of time, implying a potential decrease of the explanatory power of this variable.
For the case of the day of publication (DateDay), we use a twoway segmentation, differentiating between those posts published
on weekends than those published on weekdays, looking for potential differences in terms of interaction rates as a consequence of
the day of publication.
Table 2 shows the number of observations for each variable according to the categories abovementioned.
To proxy for content popularity in Facebook fan pages two dependent variables were chosen: Likes and Comments. Both measures have been widely used as measures of publication impact (De
Vries et al., 2012; Pletikosa Cvijikj & Michahelles, 2013). On the one
hand, a post with a high number of likes may indicate that its content
is of interest, increasing its probability to be liked by someone, and
thus, disseminating the brand message to a broader number of potential customers. On the other hand, a high number of comments
of a post also represent a kind of success or impact as it implies
having people spending their time giving their opinions and thoughts.
Given this specication, two models are therefore tested: a rst
one explaining the number of likes, and a second one aiming at shedding some light on those elements that enhance the number of comments on brand fan pages.

3.2 Data collection

3.4 Method

The data collection was gathered manually over one month, from
March 21 to April 21, 2011. In order to obtain relevant information,

The empirical analysis is based on multiple OLS linear regressions for each dependent variable, using the stepwise method with

3 Methodology
3.1 Sample

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

Fig. 2. Variables collected from Facebook.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

Table 1
Variables denition.
Id

Variable

Explanation

Codication and comments

Followers

Number of users that follow the travel agency Facebook fan page

Characters

Post length measured by the number of characters

Links

Number of links within the post

4a

Images

Number of images within the post

4b

Videos

Number of videos within the post

Time

Publication time of the post

DateDay

Publication date of the post

Likes

Number of likes that the post has got

Comments

Number of comments that the post has

Numerical 0.
Captured at the beginning of the data collection. It has been
transformed applying natural logarithm function to better t a normal
distribution and improve the explanatory power of the model.
Numerical 0.
Characters of the links are also considered.
Nominal-dichotomic
0: no links
1: 1 or more links
Nominal-dichotomic.
0: no images
1: 1 or more images
Nominal-dichotomic.
0: no videos
1: 1 or more videos
Nominal-dichotomic.
0: non-business hours (0:007:59 and 18:0023:59 on Monday to
Thursday; 0:007:59 and 15:0023:59 on Friday; Saturday and Sunday
at all hours)
1: business hours (8:0017:59 on Monday to Thursday; 8:0014:59
on Friday)
Nominal-dichotomic.
0: weekend (from Friday at 15:00 to Sunday at 23:59)
1: weekday (the remaining time)
Numerical 0.
It has been transformed using natural logarithm to better t a normal
distribution and improve the explanatory power of the model.
Numerical 0.
It has been transformed using natural logarithm to better t a normal
distribution and improve the explanatory power of the model.

the criteria Probability of F <= .050 for entering variables into the
regression model and Probability of F >= .100 for removing them.
No missing values were found. Outliers, those observations with
studentized residual that exceed 3 or +3, were eliminated. In order
to guarantee a normal distribution of the residuals, we used natural
logarithms of all the dependent variables, these being calculated as
LN(Likes+1) and LN(Comments+1). Fig. 3 indicates that the errors obtained for the different regression models are normally distributed, conrming the validity of this approach.
Before proceeding with the empirical analysis, further clarication for the Followers variable is required, as the number of followers is a characteristic of the travel agencys Facebook fan page and
is collected at the very beginning of the study, whereas the rest of
the independent variables considered are specic features of each
post. In order to improve the explanatory power of this variable we
transformed it using the natural logarithm. Therefore, the models
tested are designed with LN(Followers) instead of Followers.

Table 2
Sample size after ltering by categories of dummy variables.
Variable

Category1: Number
of elements
Likes Modela (Comments
Modelb)

Category2: Number
of elements
Likes Modela (Comments
Modelb)

Links
Images
Videos
Time
DateDay

no links: 87 (88)
no images: 113 (115)
no videos: 144 (145)
non-business hours: 25 (25)
weekend: 16 (16)

1 or more links: 75 (76)


1 or more images: 49 (49)
1 or more videos: 18 (19)
business hours: 137 (139)
weekday: 146 (148)

a
b

From a total of 162 observations after having discarded two outliers.


From a total of 164 observations.

Fig. 3. QQ Plots of Standardized Residual.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

Table 3
Results overview.
Models

Likes model

Comments model

Ra
R square
Adj. R square
Signicant variablesb
Images
Videos
Links
DateDay
Time
Characters
LN(Followers)

0.743 (high)
0.553
0.541

0.595 (moderate)
0.355
0.338

(++)
(++)

(+)
(-)
(+)

(++)
(++)

(+)

(+) / () Positive or Negative interrelationship at the level of 0.05.


a Codication for r: Very low [.0, .2); Low [.2, .4); Moderate [.4, .6); High [.6, .8);
Very High [.8, 1].
b (++) / () Positive or Negative interrelationship at the level of 0.001.

4 Results
An overview of the results is presented in Table 3, summarizing the main ndings of the two models tested: Likes model and
Comments model.
4.1 Likes model
Following the abovementioned procedure, we test all the hypotheses through an OLS linear regression with LN(Followers), Characters, Images, Links, Videos, Time and DateDay as independent
variables and LN(Likes+1) as the dependent variable for brand post
popularity. Two observations were eliminated. Table 4 provides the
full description of the coecients for the signicant variables in the
Likes model.
The explanatory power of the model (R-square) is 55.3% and the
ANOVA test calculates a value of 48.468 (p-value < 0.001) for F(4;157),
evidencing a signicant and positive linear effect (p-value < 0.001)
of certain factors (LN(followers), Characters, Images and Videos) over
the number of Likes. The formulation of the resulting model is expressed in Equation 1.

LN ( Likes + 1) = 0.929 * Videos + 0.673 * Images


+ 0.632 * LN ( Followers )
+ 0.003 * Characters 3.951 +

(1)

In order to guarantee the statistical correctness of the model, we


tested the residuals behavior in terms of normality, independence, homoscedasticity and multicollinearity assumptions. Although some authors suggest that the normal distribution of the
residuals is not a requirement of the linear regression model (Greene,
2003), we tested it to strengthen the robustness of the model. The
normal QQ plot of standardized residual (see Fig. 3) as well as the
KolmogorovSmirnov normality test (p-value = 0.200) and
the ShapiroWilk test (p-value = 0.054) indicate that we cannot refuse
the hypothesis of normality.

Table 4
Coecients for signicant variables in Likes Model.

Videos
Images
LN(Followers)
Characters
(Constant)
a
b

p-value < 0.05.


p-value < 0.001.

Std error

VIF

0.929
0.673
0.632
0.003
3.951

0.215
0.145
0.089
0.001
0.798

4.31b
4.65b
7.11b
5.09b
4.95b

1.250
1.205
1.189
1.370

Fig. 4. XY Plots of Predicted value and Residual.

The independence assumption is also accomplished according


to DurbinWatsons test which calculates a value of 1.665 within
the interval [1.5, 2.5], meaning that results appear not to be auto
correlated. Fig. 4 (see Likes model) also demonstrates that the
homoscedasticity assumption is fullled. Likewise, no collinearity
problems were observed, as the maximum VIF index calculated was
1.370 for the Characters variable (Allison, 1999; Belsey, Kuh, & Welsch,
1980).
Additionally, other assumptions of the regression model are accomplished: the expected value of the residuals is 0; there is no signicant correlation between the residuals and the independent
variables, and there are neither outlier observations nor critical values
since the standardized residuals interval is [2.109, 2.883] and the
maximum value of the Cooks distance of the residuals is lower than
1 (0.111).
All these characteristics corroborate the robustness and appropriateness of the model tested, where independent variables explain
55.3% of the Likes variable.
In terms of the results, our ndings for richness highlight the positive impact of images and videos which contributes to attract users
attention and are likely to be transformed into likes. These results
validate hypotheses H1a and H1b. However, hypothesis H1c is not
supported as there is no evidence that the inclusion of links in a post
has any effect.
Our model fails in establishing a connection between the popularity (in terms of likes) and time frame variables. As shown in
Equation 1, neither Time nor DateDay variables were entered in the

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

Table 6
Expected and obtained results by hypotheses.

Table 5
Coecients for signicant variables in Comments Model.

Images
Time
LN(Followers)
Links
(Constant)
a
b

Std error

VIF

Hypothesis

Expected

Likes Model

Comments Model

0.813
0.651
0.293
0.627
1.320

0.230
0.227
0.112
0.205
1.032

3.53a
2.87a
2.62a
3.05a
1.28

1.721
1.033
1.076
1.625

H1a (Images)
H1b (Videos)
H1c (Links)
H2a (DateDay)
H2b (Time)

(+)
(+)
(+)
(+)
(+)

Supported
Supported
Not supported
Not supported
Not supported

Supported
Not supported
Not supported (negative effect)
Not supported
Supported

p-value < 0.05.


p-value < 0.001.

model. Consequently, this prevents us to conrm hypotheses H2a


and H2b.
Looking at the control variables, we observe that, as expected,
a higher number of followers imply a higher impact on the number
of likes. Surprisingly, the number of characters is found to exert a
signicant positive inuence on the dependent variable. In Facebook, where the length of the message has no restrictions, a longer
text may suggest a post offering more detailed information.
4.2 Comments model
All the hypotheses are tested through an OLS linear regression
using the same independent variables than those used in the previous model, but having the LN(Comments+1) as the dependent variable for popularity. Neither missing values nor outliers were found.
The resulting model (Equation 2) explains 35.5% of the
variance, and the ANOVA test calculates a value of 21.828 (pvalue < 0.001) for F(4,159), conrming a moderate linear relationship between the dependent and some independent variables.

LN (Comments + 1) = 0.813 * Images + 0.651* Time


+ 0.293 * LN ( Followers )
0.627 * Links 1.320 +

(2)

As shown in Table 5, there is a positive signicant effect between


LN(Comments+1) and Images (p-value < 0.05), Time (p-value < 0.05)
and LN(Followers) (p-value < 0.05) variables. On the contrary Links
has negative effect (p-value < 0.05).
The statistical correctness of the model has also been tested following the same procedure as for the Likes model. Here, we observed that the normal QQ plot of standardized residual (see Fig. 3)
corroborates the normal distribution of the residuals. The independence assumption is also accomplished, obtaining a value of 2.115
within the interval [1.5, 2.5] in the DurbinWatsons test. Although Fig. 4 for the Comments model is not as clear as in the former,
we believe that the homoscedasticity assumption is validated as
Cameron and Trivedis IM-test shows (p-value = 0.127). Collinearity assumption is also fullled since the maximum VIF index calculated is 1.721 for Images variable (Allison, 1999; Belsey et al., 1980).
We also control for the non-existence of outliers and critical values,
obtaining an interval for standardized residuals of [2.447, 2.717]
and all Cooks distances of the residuals being lower than 1. Additional analyses (the expected value of the residuals is 0 and there
is no signicant correlation of the residuals and independent variables) further corroborate the validity of the model despite its moderate power of explanation.
For richness, our results prove that images help increase the
number of comments a post gets, supporting hypothesis H1a.
However, there is no evidence that videos inuence the number of
comments, meaning that hypothesis H1b is not supported. Contrary to what we expected, it is shown that links have a negative
effect on post popularity in terms of comments, therefore hypothesis H1c is rejected.
Regarding the variables representing the time frame, our ndings indicate that the hour of publication (Time) also plays a key role.

Particularly, our results corroborate the hypothesis that those posts


published during business hours are more likely to be commented
than those published outside this schedule. Consequently, we can
assume that H2b is supported. Nevertheless, the effect of the day
of the week dilutes as this variable does not enter the model specications with a signicant coecient, signaling that H2a is not
supported.
As for the control variables, we nd that the variable Followers
exhibits a similar behavior as in the Likes model, meaning that having
a large amount of followers positively inuences the number of comments a post may get, indicating that more people is expected to
have access to it.
An overview of nals results through hypotheses testing is shown
in Table 6.
5 Discussion and future research avenues
In this paper, we have analyzed hard criterion factors that inuence the popularity of brand posts published on Facebook and
tested them for the Spanish travel agency sector. These factors are
not related to the meaning of the content but represent structural
characteristics of posts.
Following the conceptual model, the structural characteristics of
a post have been classied according whether they refer to the vividness of its content (richness) or indicate time frame (time and date
of publication).
With respect to the richness, our results point towards the importance of the use of images which are proven to cause the greatest level of engagement, attracting more easily users attention and
turning this attention into likes and comments. This result is in accordance with previous studies, suggesting that images are an important element of the posting strategy which signicantly increase
brand post popularity.
Creative endeavors in the form of videos to enrich the content
of a post only apply when post popularity is measured through the
number of likes. In our interpretation, this result could signal that
images are easier to digest and in a few seconds users can write a
short comment about the feelings/opinions that the picture has
invoked on them. However, the process of commenting requires users
to dedicate more time to rst assimilate the content and second to
publicly assess it by writing an opinion. Undoubtedly, commenting requires an additional effort in comparison with liking (only one
click is needed).
Results also show that links are negatively inuencing
the number of comments. When publishing a link, Facebook shows
a small summary of the content of the destination page. This outline
may be evocative enough to motivate likes (although no relationship has been found), whereas to be able to comment users need
to visit an external page and consume its content. Nevertheless, clicking on the link implies navigating away from Facebook to the destination page, increasing the risk of users not coming back and
commenting.
At this point, it worth noting that the choice of avoiding a priori
judgments about progressive levels of richness (low, medium and
high) has been effective. Operating as in this paper, we have been

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

able to individually investigate different behaviors according to


content type. In this respect, our ndings support our argument that
the use of different types of richness does not necessarily drive to
the same effect on brand post popularity.
Regarding the potential effect that time frame factors have over
the popularity of a post, our ndings are limited. We only found evidence to support the hypothesis that relates the time of publication (differentiating between business and non-business hours) and
the number of comments. This result suggests that people get notice
of new posts during business hours because they are connected and
probably, are in front of a computer. This could also be related with
the use of different devices. Writing a comment using desktop
devices is undoubtedly easier than doing it from mobile ones. It is
supposed that outside business hours more users are connecting
to Facebook through their mobile terminals which could make difcult to write comments. Although data gathered do not include this
sort of information, if posts are published during business hours they
are more likely to be commented. On the contrary, liking activity
is not inuenced by posting time. No effect occurred in regard to
the day of the week (weekdays vs weekends), neither for likes nor
for comments.
Finally, the control variables also provide remarkable information. When the content published in the fan page (by the brand or
even by fans) meets the characteristics to become popular, it is virally
disseminated through the network of fans, fans friends, friends of
friends, and so on. Consequently, the larger the number of followers, the easier it will be for the company to spread their message
and reach sizeable audiences.
Concerning the length of the post, our results support
the convenience of writing larger posts for increasing the number
of likes. On the contrary, the length of the message is not signicant for the number of comments. This nding contrasts with the
one reported by Buddy Media Inc. (2011) that posts shorter than
80 characters have, on average, 27% higher engagement. This disagreement could be grounded in very different causes such as methodological differences (the report denes engagement as a
combination of likes and comments), cultural differences between
Spanish and north American audiences (those used in the
aforementioned report), the language idiosyncrasy or industry
specicities.
To conclude, our ndings reinforce the abovementioned notion
that liking and commenting actions are of different nature and need
to be examined separately. Not all determinants which are found
to positively enhance one form of brand post popularity (likes/
comments) also have a positive effect on other forms (comments/
likes). Writing a comment is a much more time-consuming process
than liking and is related to different motivations triggered
by the meaning of the content. For instance, a short question like
What are your plans for this weekend? seems to be more
likely to motivate comments than likes. People will comment when
the content is really meaningful for them or request them to act.
Writing a comment seems more dependent on the emotions and
feelings. Also users may be willing to comment if they perceive
specic benets (e.g. discounts, special offers) or when the content
of the post causes an emotional impact or a feeling that overwhelms the reader (Hettler, 2010). Given the increasing importance that SNS are gaining as marketing tools and in the light of
our ndings, we believe that further research efforts in this direction are necessary.
Given the aforementioned considerations, a recommendation for
further studies relates to the effect that soft criteria factors have over
likes and comments, as this is one of the main limitations of the
present work. More sophisticated models of brand post popularity can be developed by including both criteria (soft and hard). These
models can be enhanced through structural equation modelling
(SEM) to benet from the ability to construct latent variables and

to reect indirect casual relationships that may arise between factors.


Also the use of text mining and sentiment mining methods such as
those reported by Barbier and Liu (2011) and Aggarwal and Wang
(2011) would enable capturing bigger data samples and the incorporation of new variables in the models tested. Notwithstanding,
we are aware of the diculties in obtaining reliable data, as there
are some concerns about the consistency of these automatic algorithms when capturing and analyzing the meaning of content, drawn
or recorded in posts, and the potential emotions that may arise
among users.
Indeed, a model using variables following both the soft and the
hard criteria may lead to models with the highest predictive and
explanatory power, shedding some light on those factors that help
rms to engage more eciently with their customers, improving
their current communication channels. Also, investigation comparing results in different SNS could reveal interesting facts for
marketers.

6 Conclusions
This work empirically contributes to a better understanding of
the use of social media marketing strategies. Particularly, we have
identied those structural factors of posts published on Facebook
brand pages that are observed to inuence brand post popularity,
measured through the number of likes and comments. To do this,
we have focused on a sample of Spanish travel agencies with a Facebook fan page.
Results obtained point to some guidelines for improving the liking
of posts published on Facebook brand pages. Community managers should include images and videos which seem to better attract
customers attention, especially in the case of images. As for the
length of the post, moderators should not be worried about writing
to many characters if this is essential for a good understanding of
the content. In this sense, we found that the number of characters
employed is positively correlated with the number of likes. Finally,
the low statistical signicance of time frame factors prevents us from
formulating recommendations related with the day and the time
that best helps increase the number of likes.
Guidelines for improving the number of comments differ from
those suggested to increase the number of likes. In this case, community managers should look for posts that include images because
this is the only richness factor that is positively related with the
number of comments. Furthermore, the avoidance of links would
report more comments. Links can act as barriers, driving users to
external websites that make them forget returning to the Facebook fan page and leave a comment. Another advice that moderators should follow consists in publishing during business hours as
it seems to improve users willingness to comment. However, this
strategy should be taken with a grain of salt because as shown in
previous studies there is some controversy in the exact denition
of the most effective time period.
Another remarkable conclusion emerging from this study is that
images are more powerful than videos in increasing consumers engagement. The variable Image is positively signicant in both models
(likes and comments) whereas Videos is only signicant when popularity is expressed in terms of likes.
It is expected that both the conceptual model and the results
obtained through this empirical analysis provide meaningful
theoretical and managerial implications for rms and marketers,
and especially for those professionals working on the travel agency
sector. Acknowledging the effects that structural characteristics of
posts have on users involvement (in terms of likes and comments) may help community managers to effectively exploit social
networking sites within the integrated marketing communications of the brand.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
10

F. Sabate et al./European Management Journal (2014)

References
aDigital. (2011). Uso de Facebook por parte de las empresas espaolas. Asociacin
Espaola de la Economa Digital. <http://www.adigital.org/servicios/uso-defacebook-por-parte-de-las-empresas-espanolas> Accessed 05.21.13.
Aggarwal, C. C., & Wang, H. (2011) Text mining in social networks. In C. C. Aggarwal
(Ed.), Social network data analytics (pp. 353378). Springer US, Hawthorned, NY.
<http://www.springerlink.com/content/n612222780l37460/abstract/> Accessed
10.10.12.
Agresta, S., & Bough, B. B. (2011) Perspectives on social media marketing: the agency
perspective, the brand perspective. Course Technology, Boston, MA.
Allison, P. D. (1999) Logistic regression using the SAS system theory and application.
SAS Institute, Cary (N.C.).
Baltas, G. (2003) Determinants of Internet advertising effectiveness: an empirical
study. International Journal of Market Research 45(4), 505513.
Barbier, G., & Liu, H. (2011) Data mining in social media. In C. C. Aggarwal (Ed.), Social
network data analytics (pp. 327352). Springer US. <http://www.springerlink
.com/content/v2h43358124qu862/abstract/> Accessed 10.10.12.
Belsey, D. A., Kuh, E., & Welsch, R. E. (1980) Regression diagnostics: identifying inuential
data and sources of collinearity. John Wiley, New York, NY.
Blanke, J., & Chiesa, T. (2012). Travel & tourism competitiveness report 2011.
Beyond the downturn. <http://www3.weforum.org/docs/WEF_TravelTourism
Competitiveness_Report_2011.pdf> Accessed 05.21.13.
Brookes, E. J. (2010). The anatomy of a facebook post. Study on post performance by
type, day of week, and time of day. Vitrue-Oracle. <http://www.vitrue.com/wpcontent/themes/Vitrue-3.0/white-papers/anatomy_of_fb_wp.pdf> Accessed
09.05.11.
Brown, J., Broderick, A. J., & Lee, N. (2007) Word of mouth communication within
online communities: conceptualizing the online social network. Journal of
Interactive Marketing 21, 220. doi:10.1002/dir.20082.
Buddy Media Inc. (2011). Strategies for effective facebook wall posts: a statistical review.
<http://www.buddymedia.com/newsroom/2011/04/introducing-our-latest
-research-strategies-for-effective-facebook-wall-posts-a-statistical-review/>
Accessed 07.30.11.
Bughin, J., & Manyika, J. (2009). How businesses are using Web 2.0: a McKinsey Global
Survey. McKinsey Quarterly, (09). <http://www.mckinsey.com/insights/
business_technology/how_companies_are_beneting_from_web_20
_mckinsey_global_survey_results> Accessed 06.17.13.
Buhalis, D., & Law, R. (2008) Progress in information technology and tourism
management: 20 years on and 10 years after the internet the state of eTourism
research. Tourism Management 29, 609623. doi:10.1016/j.tourman.2008.01
.005.
Chevalier, J. A., & Mayzlin, D. (2006) The effect of word of mouth on sales: online
book reviews. Journal of Marketing Research 43, 345354. doi:10.1509/
jmkr.43.3.345.
Constantinides, E., & Fountain, S. J. (2008) Web 2.0: conceptual foundations and
marketing issues. Journal of Direct, Data and Digital Marketing Practice 9, 231244.
doi:10.1057/palgrave.dddmp.4350098.
Daft, R. L., & Lengel, R. H. (1986) Organizational information requirements, media
richness and structural design. Management Science 32, 554571. doi:10.1287/
mnsc.32.5.554.
De Vries, L., Gensler, S., & Leeang, P. S. H. (2012) Popularity of brand posts on brand
fan pages: an investigation of the effects of social media marketing. Journal of
Interactive Marketing 26, 8391. doi:10.1016/j.intmar.2012.01.003.
Dhar, V., & Chang, E. A. (2009) Does chatter matter? The impact of user-generated
content on music sales. Journal of Interactive Marketing 23, 300307. doi:10.1016/
j.intmar.2009.07.004.
Dholakia, U. M., Bagozzi, R. P., & Pearo, L. K. (2004) A social inuence model of
consumer participation in network- and small-group-based virtual communities.
International Journal of Research in Marketing 21(3), 241263. doi:10.1016/
j.ijresmar.2003.12.004.
Dholakia, U. M., & Durham, E. (2010) One Cafe Chains Facebook experiment. Harvard
Business Review 88, 26.
Duan, W., Gu, B., & Whinston, A. B. (2008) The dynamics of online word-of-mouth
and product sales an empirical investigation of the movie industry. Journal of
Retailing 84, 233242. doi:10.1016/j.jretai.2008.04.005.
Fischer, E., & Reuber, A. R. (2011) Social interaction via new social media: (How) can
interactions on Twitter affect effectual thinking and behavior? Journal of Business
Venturing 26, 118. doi:10.1016/j.jbusvent.2010.09.002.
Fortin, D. R., & Dholakia, R. R. (2005) Interactivity and vividness effects on social
presence and involvement with a web-based advertisement. Journal of Business
Research 58, 387396. doi:10.1016/S0148-2963(03)00106-1.
Fournier, S., & Avery, J. (2011) The uninvited brand. Business Horizons 54, 193207.
doi:10.1016/j.bushor.2011.01.001.
Golder, S. A., Wilkinson, D. M., & Huberman, B. A. (2007) Rhythms of social interaction:
messaging within a massive online network. In C. Steineld, B. T. Pentland, M.
Ackerman, & N. Contractor (Eds.), Communities and technologies (pp. 4166).
Springer, London. <http://link.springer.com/chapter/10.1007/978-1-84628-905
-7_3> Accessed 04.01.14.
Greene, W. H. (2003) Econometric analysis. Prentice Hall, Upper Saddle River, NJ.
Hanna, R., Rohm, A., & Crittenden, V. L. (2011) Were all connected: the power of the
social media ecosystem. Business Horizons 54, 265273. doi:10.1016/
j.bushor.2011.01.007.
Hansen, D. L., Schneiderman, B., & Smith, M. A. (2011) Analyzing social media networks
with NodeXL: insights from a connected world. M. Kaufmann, Amsterdam; Boston.

Hennig-Thurau, T., Malthouse, E. C., Friege, C., Gensler, S., Lobschat, L., Rangaswamy,
A., et al. (2010) The impact of new media on customer relationships. Journal of
Service Research 13(3), 311330. doi:10.1177/1094670510375460.
Hettler, U. (2010) Social media marketing: marketing mit Blogs, sozialen Netzwerken
und weiteren Anwendungen des Web 2.0. Oldenbourg, Mnchen.
Heymann-Reder, D. (2011) Social media marketing: strategien fr Sie und
Ihr Unternehmen. Addison Wesley in Pearson Education Deutschland,
Mnchen.
Hollenbeck, C. R., & Kaikati, A. M. (2012) Consumers use of brands to reect their
actual and ideal selves on Facebook. International Journal of Research in Marketing
29(4), 395405. doi:10.1016/j.ijresmar.2012.06.002.
Hong, L., Dan, O., & Davison, B. D. (2011) Predicting popular messages in Twitter. In
Proceedings of the 20th International Conference Companion on World Wide Web
(pp. 5758). ACM., New York, NY, USA. doi:10.1145/1963192.1963222.
IAB Spain Research & Elogia. (2011). III estudio sobre redes sociales en internet. IAB
Spain Research and Elogia. <http://www.iabspain.net/wp-content/plugins/
download-monitor/download.php?id=73> Accessed 05.21.13.
Kilian, T., & Langner, S. (2010) Online-Kommunikation: Kunden zielsicher verfhren und
beeinussen. Gabler, Wiesbaden.
Kozinets, R. V., de Valck, K., Wojnicki, A. C., & Wilner, S. J. S. (2010) Networked
narratives: understanding word-of-mouth marketing in online communities.
Journal of Marketing 74, 7189.
Lawrence, D., Pownal, C., Jrg, D., & Carmo, C. (2011). 2011 Fortune Global 100 social
media study. The Burson-Marsteller Blog, February 15. <http://www.burson
-marsteller.com/Innovation_and_insights/blogs_and_podcasts/BM_Blog/Lists/
Posts/Post.aspx?ID=254> Accessed 05.21.13.
Mangold, W. G., & Faulds, D. J. (2009) Social media: the new hybrid element of the
promotion mix. Business Horizons 52, 357365. doi:10.1016/j.bushor.2009.03
.002.
Moldovan, S., Goldenberg, J., & Chattopadhyay, A. (2011) The different roles of product
originality and usefulness in generating word-of-mouth. International Journal of
Research in Marketing 28(2), 109119. doi:10.1016/j.ijresmar.2010.11.003.
Pletikosa Cvijikj, I., & Michahelles, F. (2011) A case study of the effects of moderator
posts within a Facebook brand page. Social Informatics 6984, 161170.
Pletikosa Cvijikj, I., & Michahelles, F. (2013) Online engagement factors on Facebook
brand pages. Social Network Analysis and Mining 3, 843861. doi:10.1007/s13278013-0098-8.
Radwanick, S., Lipsman, A., & Aquino, C. (2011). Its a social world: top 10 need-to-knows
about social networking and where its headed. comScore. <http://
www.comscore.com/Press_Events/Presentations_Whitepapers/2011/it_is
_a_social_world_top_10_need-to-knows_about_social_networking> Accessed
05.21.13.
Rehmani, M., & Khan, M. I. (2011) The impact of E-media on customer purchase
intention. International Journal of Advanced Computer Science and Applications 2,
100103.
Rutz, O. J., & Bucklin, R. E. (2011) From generic to branded: a model of spillover in
paid search advertising. Journal of Marketing Research 48, 87102. doi:10.1509/
jmkr.48.1.87.
Sabate, F., Berbegal, J., Consolacin, C., & Caabate, A. (2009) SEO strategies in
booksellers sector. Intangible Capital 5, 321. doi:10.3926/ic.2009.v5n3.p321
-346.
Sabate, F., Canabate, A., Velarde-Iturralde, M.-A., & Grinon-Barcelo, R. (2010) Use of
internet promotion strategies by the Spanish travel agencies. Profesional De La
Informacion 19, 149159. doi:10.3145/epi.2010.mar.05.
Sashi, C. M. (2012) Customer engagement, buyer-seller relationships, and social media.
Management Decision 50(12), 253272. doi:10.1108/00251741211203551.
Scott, D. M. (2007) The new rules of marketing and PR: how to use news releases, blogs,
podcasting, viral marketing, & online media to reach buyers directly. John Wiley &
Sons, Hoboken, NJ.
Sierra Snchez, J. (2012) Factors inuencing a students decision to pursue a
communications degree in Spain. Intangible Capital 8, 4360. doi:10.3926/ic.277.
Silverman, D. (2012). IAB internet advertising revenue report. 2011 Full Year results.
Interactive Advertising Bureau and PricewaterhouseCoopers. <http://
www.iab.net/media/le/IAB_Internet_Advertising_Revenue_Report_FY
_2011.pdf> Accessed 05.21.13.
Simmons, G. (2008) Marketing to postmodern consumers: introducing the internet
chameleon. European Journal of Marketing 42, 299310. doi:10.1108/
03090560810852940.
Singh, V. K., Jain, R., & Kankanhalli, M. (2011) Mechanism design for incentivizing
social media contributions. In S. C. H. Hoi, J. Luo, S. Boll, D. Xu, R. Jin, & I. King
(Eds.), Social media modeling and computing (pp. 121143). Springer London,
London. <http://rd.springer.com/chapter/10.1007/978-0-85729-436-4_6> Accessed
10.10.12.
Smith, A. N., Fischer, E., & Yongjian, C. (2012) How does brand-related user-generated
content differ across YouTube, Facebook, and Twitter? Journal of Interactive
Marketing 26, 102113. doi:10.1016/j.intmar.2012.01.002.
Smith, P. R., & Zook, Z. (2011) Marketing communications: integrating oine and online
with social media. Kogan Page, London [u.a.].
Sterne, J. (2010) Social media metrics: how to measure and optimize your marketing
investment. John Wiley, Hoboken, N.J.
Suh, B., Hong, L., Pirolli, P., & Chi, E. H. (2010). Want to be retweeted? Large scale
analytics on factors impacting retweet in Twitter network. In Social Computing
(SocialCom) 2010 IEEE Second International Conference on Social Computing (pp.
177184). Presented at the 2010 IEEE Second International Conference on Social
Computing (SocialCom). Minneapolis, MN: IEEE. doi:10.1109/SocialCom.2010
.33.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001

ARTICLE IN PRESS
F. Sabate et al./European Management Journal (2014)

Sun, E., Rosenn, I., Marlow, C., & Lento, T. (2009). Gesundheit! Modeling contagion
through Facebook news feed. In Proceedings of the third international AAAI
conference on weblogs and social media. Presented at the International AAAI
Conference on Weblogs and Social Media, San Jose, CA: AAAI Press. <http://
snap.stanford.edu/class/cs224w-readings/sun09contagion.pdf> Accessed 10.10.12.
Trusov, M., Bucklin, R. E., & Pauwels, K. (2009) Effects of word-of-mouth versus
traditional marketing: ndings from an internet social networking site. Journal
of Marketing 73, 90102.
Tuten, T. L. (2008) Advertising 2.0: social media marketing in a Web 2.0 world. Praeger
Publishers, Westport, CT.
Urban, G. L. (2003). Customer advocacy: is it for you? (Working Paper No. 175).
<http://digital.mit.edu/research/papers/175_Urban_Trust.pdf> Accessed 10.10.12.

11

Verhoef, P. C., & Lemon, K. N. (2013) Successful customer value management: key
lessons and emerging trends. European Management Journal 31(1), 115.
doi:10.1016/j.emj.2012.08.001.
Xiang, Z., & Gretzel, U. (2010) Role of social media in online travel information
search. Tourism Management 31, 179188. doi:10.1016/j.tourman.2009.02
.016.
Ye, Q., Law, R., & Gu, B. (2009) The impact of online user reviews on hotel room sales.
International Journal of Hospitality Management 28, 180182. doi:10.1016/
j.ijhm.2008.06.011.
Zhang, L., Peng, T.-Q., Zhang, Y.-P., Wang, X.-H., & Zhu, J. J. H. (2014) Content or context:
which matters more in information processing on microblogging sites. Computers
in Human Behavior 31, 242249. doi:10.1016/j.chb.2013.10.031.

Please cite this article in press as: Ferran Sabate, Jasmina Berbegal-Mirabent, Antonio Caabate, Philipp R. Lebherz, Factors influencing popularity of branded content in
Facebook fan pages, European Management Journal (2014), doi: 10.1016/j.emj.2014.05.001